Could this be the beginning of the end of AI hype?
4 Likes
Not all AI applications are the same. What matters is having the cash flow to amortise the investment. The Global Crossing bankruptcy is a good example. They had a great plan but unexpectedly what they could charge dropped like a lead balloon and they could not pay off their debt.
What are you going to do with your AI investment?
- OpenAI has to find customers for its AI services.
- Tesla uses their AI in-house
Which is more likey to go broke?
The Captain
3 Likes
Itâs been a long time coming, but yes.
5 Likes
And Alphabet (Google), Microsoft, and Meta (Facebook) have announced EVEN HIGHER capex numbers yesterday. Itâs crazy large numbers they are stating.
Thereâs surely going to be, and is already, a boom for the semiconductor companies - TSMC, Nvidia, AMD, support companies, memory companies like Micron, etc. But then there will be the eventual bust as well. Anyone who can time it reasonably accurately will see outsized returns.
3 Likes
Maybe. AI is definitely seeing some headwinds. Inflation, lawsuits, and regulatory interventions will definitely have an impact on the bottom line. That said, itâs probably still a bit early to bet against them.
It is crazy, but theyâre not all the same. Meta is dumping money in, but doesnât have the cloud computing capability like the others. That makes it more difficult for them to profit off of their capex investment. Todayâs slide reflects that reality. It also doesnât help when the CFO says stuff like -
âWe arenât providing a specific outlook for 2027 capex. And we are, frankly, undergoing a very dynamic planning process ourselves as weâre working through what our capacity needs will be over the coming years,â Meta CFO Susan Li told analysts. âOur experience so far has been that we have continued to underestimate our compute needs even as we have been ramping capacity significantly as the advances in AI have continued and our teams continue to identify compelling new projects and initiatives.â
âSo weâre going to continue building out our infrastructure with flexibility in mind,â she added. âAnd if we end up not needing as much as we anticipate, we can choose to bring it online more slowly or reduce our spending in future years as we grow into the capacity that weâre building now.â
2 Likes
Itâs not just the crazy size (and uncertainty) of the capex spend. Itâs the basic question of why are they doing this? This isnât their business model. Theyâre not even setting themselves up to be selling this AI product to other users or companies, but using it themselves. But they are not a software development company, they are not a tech company - they are an ADVERTISING company.
Unlike Microsoft, Amazon, & Google, Meta does not run a datacenter business. When the other 3 companies spend tens of billions on AI Capex, they immediately turn around and sell usage of that computing capacity to other companies. In other words, they get immediate revenue from their investment. Meta, on the other hand, only has the expense line.
[SNIP]
Contrast the path that Meta has chosen with the one that Apple has chosen. Apple, like Meta, does not operate a datacenter business. Their business model is that they sell hardware and services to consumers at a profit. Rather than spend tens of billions of dollars to chase a proprietary AI model, they have announced that they will be using the Gemini engine from Google to power their next generation Siri app. They will pay Google $1 billion per year to develop a custom Gemini model. That is a mature management decision which indicates that they understand their own business model.
Well, I think I have beaten this horse enough. You get the point. Of all the big spenders on AI, it is Meta that makes me shake my head the most.
Meta should be a buyer of AI, not a developer | by Michael Liss | Medium
5 Likes
I agree that Meta is an advertising company. But perhaps Google is one too - 77.8% of Googleâs revenue comes from advertising.
I think Google feels that advertising may have peaked (at least as a percentage of GDP) so they want to diversify into providing cloud/AI services.
1 Like
I am using Claude professionally as Pro 20x user, $200/month. My productivity is way up. It is worth far more than the $200 monthly.
That said larger enterprise rates to corporations are not such a grand deal.
AI greatly enhances the one-man shop. The fellow who started his own GLP-1 marketing and distribution firm is doing $1.8 billion in the last twelve months. He made his brother his only other partner/employee.
1 Like
All our emails are now AI. Meaning the materials are recreated to draw you into viewing them. This opens up a new art in marketing to create greater interest in your spam.
The invitation to Google is to have people use it like YouTube for hours at a time. Right now, the search is a limited time window. People find what they want and leave. Now you can get all sorts of answers while engaged in the real world. The time spent on Google search may be increasing. When I search for an answer, I skip the bloggers; I go to Google AI because a factoid is all I want.
Cloud/AI is growing much faster than the ad business, but the ad revenue still grew by 11% YoY last quarter. Not shabby at all.
Cloud was almost off charts. Revenues were up 63% and margins grew from 18% to 33%.
4 Likes